On 3 Apr, 22:51, "David Kane" wrote:
"Mark Houlsby" wrote in message

oups.com...





On 3 Apr, 21:31, "David Kane" wrote:

nor that drawing two games against a GM opponent are less of an
achievement [for me!] than a win and a loss,

Mathematically, it *is* less of an achievement
because GMs don't lose very often.

Chessically drawing is better.

How you feel
about the achievement is, of course, up to you.

nor that limited
Serious Money in chess is proof that something is Wrong,

Proof, possibly not. But it sure has to be evidence of something.
Why deny it?

Chess is too complex. End of story. Why worry about it?

Go is *more* complex and less accessible than chess,
yet appears to have a much healthier professional game.


Yes. Maybe it's because its adherents are more sophisticated than
chess fans who *demand* poor-quality games.

Some relevant factors:
Go doesn't have draws.

Maybe that's another reason. In fact, probably it is.

Chess not only has draws, but
has people like you who worship draws at all costs.


Ummm... I don't do that. I do recognise that draws are a necessary
part of professional chess. Also, some draws are more beautiful than
many decisive games which have won brilliancy prizes.

Go refines its rules to improve its competitions.
For example, the scoring has changed to increase
the amount of compensation given to the 2nd player.
The chess world clings to demonstrably idiotic
scoring practices and preaches to us that we
should learn to appreciate the monstrosities
which result from them.


It doesn't preach. Go watch Go.

Frankly I don't think your type will be happy until
someone wins a prize of $2243 for winning
the chess world championship, thereby falling
behind checkers.

Frankly, you're a moron.

"Dr A. N. Walker" wrote in message
...
In article ,
David Kane wrote:

Almost
every other popular activity has well-known stars
whose appeal extends well beyond that of expert
practitioners of the sport.
[...]
I'm not quite sure I get your point. I guess I'm saying that
chess has a low (marketable product)/participant ratio.

But that's not what you *were* saying. I merely pointed
you at a fairly long list of popular activities that do not have
"well-known stars", and then at the common theme of those that do
-- viz TV exposure.


But your list was full of things that are obviously nowhere
near as popular as chess, so it wasn't very persuasive.
In the US, a country with relatively little chess tradition,
the percentage of people with some interest in the game is
in double digit percentages. Ice dancing?

Put chess on reasonably prime-time TV, and
it will have stars; toss it to 4am on the minor channels and to
the back pages of the "quality" press, and it won't. Having a
"marketable product" is yet another quite different concept. ...

Your TV point was also unclear. Whether something makes it
on TV depends on how interesting it is. I'm proposing a way
that chess could be more interesting than it is.


As you seem to be aware, other activities which don't
seem any more inherently interesting than chess seem to find
a market. But chess refuses to.

... "Refuses"? Look, it's not a spectator sport. So
its "market" is elsewhe books, coaching, sponsorship, sets,
product placement, etc. You can get sponsorship not because
it's "exciting" but because it is associated with intellect,
because it's "educational", because it's something that people
who are disabled, poor, immigrant, young, old, ... can do,
because [broadly] it's *not* associated with crime, drugs, sex,
betting, .... There is quite a lot of money swilling around
chess, one way or another -- enough, world-wide, to support at
least several thousand full-time professionals [playing, coaching,
writing, directing, ...] -- it just isn't channelled into a
dozen or so big names who are always on our screens and thus
known to the general public. [But Kramnik isn't exactly in
poverty ....]

I'd love to hear your expert commentary of the last
round of the Tal Memorial tournament I previously
alluded to. If you can make that interesting, you should
be in the entertainment business.

I *am* in "the entertainment business"; and making a
game of chess interesting is trivial compared with making some
of the stuff I have to teach interesting. But no matter.

[...]
First, I don't understand your claim that computers are
one-dimensional.

You previously said you did!

They evaluate many different factors and then
combine them, just like humans. The difference is that
computers will give you a concrete explanation of how
they arrive at a move,

Not in any useful way. Note that the evaluation of a
*move* is merely the result of backing the evaluations of a
huge collection of *positions* up the game tree. The factors
that went into the static evaluation of *one* position perhaps
20 or 30 ply [with extensions] away from the *current* position
will usually make little sense. "You are 0.67 pawns up because
with what I think at the moment is best play you will arrive in
15 moves at a position where your opponent has doubled pawns.
If we play *your* move, you are only 0.65 pawns up, because in
12 moves we reach a position where you have won a piece for two
pawns and he has the two bishops."

But that's a precise description of how the computer arrives at its
answer. Humans say things like "castling king-side looked risky"

whereas we humans give vague
answers.

We could give the same answers, more-or-less ["I won
a piece for two pawns and some compensation in preference to
merely keeping an edge by doubling his pawns"]. That isn't
the real difference -- which is that the human has *reasons*
[possibly mistaken] for choosing one over the other, not just
a list of numbers from which to choose the largest.

But it doesn't change the underlying nature
of the process. I don't see any reason why computers couldn't
be programmed to do other things that humans do, like reduce
counterplay, [...].

Because you need to judge the counterplay as part of
the static evaluation [already almost a self-contradiction]

I disagree. A position that is mate has no counterplay.
A position that has a lopsided material balance with
low levels of material has little counterplay, etc.
What you'll find is that a computer evaluation function will
include many things that humans describe as relating to
counterplay. The big difference is that computers evaluate
*positions* but humans include external factors. I play
1.e4 against Kramnik and he plays 1. ...f6 and offers me
a draw, I take it. Not because of my objective evaluation
of the position on the chessboard, but because I evaluate
accepting a draw as being 0.5 points better than declining
the draw and playing something else.

and then back it up the tree, which mucks up the pruning that
enables the computer to go so deep [alpha-beta more-or-less
doubles the search depth, equivalent to an exponential increase
in the speed of your computer as it searches deeper].
are partial ways around this, but it's not easy. [This is
part of the reason why humans find it hard to adjust when the
psychology of a position changes -- if you thought you were
winning, and then find you aren't, it's very hard to take a
fresh objective look at the position.]

My reason for bringing up computers' low
draw rate is that it refutes the theory that high draw rates
are inherent result of a high Elo (i.e. the perfectly played
game is a draw and humans play perfectly enough) It
suggests that the human draw rate stems from factors
external to the game [for example, overcounting draws].

Or from factors related to differences in the way
humans and computers play chess ....

Yes, but what are those specifically? Humans see
that draws help their tournament situation and play
drawish moves.

I'm just saying let's experiment with modifying the
incentives to produce more interesting competitions.
Well, that's fine. But you do your case no good by your
use of language. *I* don't feel that my own draws are not real
contests,
I never claimed that your games weren't real contests. Only
that everybody's games could be more contested if
the external incentives were different.

So you claim that I should feel that, though my own
games are real, other people are playing less competitively?
Even if true, not good politics.

I'm suggesting that in general people (including you) play
less competitively than they could but never
claimed that that makes your games "unreal".
You remind of the people who claim that
advertising doesn't work on them. But we
know that in aggregate advertising does work. Are
you claiming that your play is completely uninfluenced
by the knowledge that a drawn result will score you
points? If you played in a tournament scored with
BAP, you'd play *exactly* as you would without it?
Fact is, when strong players did play in a BAP
tournament, their behavior *was* different. They
played more interesting chess!


nor that chess is boring because some games are drawn,
I definitely never claimed that. I've been talking all along
about *excessive* draw rates.

But you divert all the time from "excessive" [if that
is true] draws to lack of competitiveness, boredom and other
consequences for popularity, marketability, ....

It's not a diversion, but a progression. Chess uses a
scoring system that makes the risk/reward of playing for a win
very poor. At the GM level, this leads to lots of draws,
and lots of tournaments with prominent strategies based
on draws. These are completely at odds with what human
beings normally consider interesting and dramatic, so few,
even the millions who enjoy playing chess, pay attention
to the games.


nor that drawing two games against a GM opponent are less of an
achievement [for me!] than a win and a loss,
Mathematically, it *is* less of an achievement
because GMs don't lose very often. How you feel
about the achievement is, of course, up to you.

But I am a good enough player to draw, occasionally,
against much stronger players, even GMs, on *merit*, without
their making a manifest error. I am not good enough to beat
a GM unless he does something really daft [which, of course,
even GMs do on occasion]. [The last time I beat a player
rated over 2400, it was a league match, and my opponent was
the only person in the room who did not know Wurzberger's
trap. Both my team and his were falling about with laughter,
and I was pleased enough to win, but didn't feel it was that
much of an achievement.]

I have similar feelings from my own games, though I'm a class
player. But it would never occur to me to argue that pride in
my draws with masters would be a valid reason to favor a
system that *obviously* harms the GM game. And it *is*
obvious - just look at the last round of the Tal Memorial
tournament I referred to.


nor that limited
Serious Money in chess is proof that something is Wrong,
Proof, possibly not. But it sure has to be evidence of something.
Why deny it?

I don't deny it. But evidence of something is not
evidence that something is Wrong.

nor
that the current Laws are outmoded.
I believe that the *scoring system* is outmoded.

Fine; you're entitled to your belief, and to try
to convince the rest of us. But the rest of us are more
likely to be convinced by rational argument that by being
told that we -- who have *chosen* to be interested in chess
as it is -- are playing by a boring, outmoded system.

I *have* wondered whether many people are attracted
to chess precisely *because* draws feature so prominently
in the game. Perhaps they fufill some psychological yearning
for "fairness". In that case, giving chess the features
we usually associate with competitions (winners, drama)
might make the game more popular in general, but less
popular to its current adherents. I think that is why so
many of the anti-anti-draw posts tend to the visceral
rather than the rational. People *like* their draws and
really don't want to analyze the consequences of their
preference.

On 3 Apr, 23:07, "Chess One" wrote:
"David Kane" wrote in message

...







"Dr A. N. Walker" wrote in message
...
In article ,
David Kane wrote:
[...] A win secured by an "unforced" blunder from a level
position is no more interesting than an agreed draw in that same
position; a draw secured by dogged defence that just succeeds is
no less interesting than a loss after dogged defence that just
fails.
Yet when people vote for "best" games, they
overwhelmingly vote for decisive ones.

That's a different matter. Firstly, if you're inviting
"votes", then these are likely to go to manifest "brilliancy"
or "drama" rather than the actual quality of the play. Secondly,
"best game" implies that its perpetrator played well; but a
*drawing* brilliancy implies that its perpetrator had played
sufficiently badly to be in a position where the rescue was
necessary.

Why is it a different matter? It's an exact expression
of the kind of chess people like. If the well-crafted draws
were as appealing as you think it is, they'd be getting votes.

It is more complex, Mark.

Ummm.... you're replying to *David*, Phil. I *know* that it's more
complex. I argued as much in RGC* a few years back.

Whatever vox populi there is is an expression of
what people think on the issue, but resolution depends on the value of this
thinking.


You imply that such thinking may be valueless? If so, then you may
have a point....

What we have here is the infamous evaluation paradigm, of material to other
factors, such as initiative. It is a thorny subject among chess players, and
even more prickly to computer folk who have no ontological basis to resolve
what either computer or people resolve as 'true'.


Well, as far as machines go, they seem to be becoming sufficiently
good at solving problems tactically that it's not impossible that--for
a while--chess may seem more closely to resemble xianqi

...

Not just physical activities either -- Scrabble, backgammon,
Othello, cruciverbalism, sudoku, philately, ... are all popular with
no "personalities" known to the UK public. Poker is different --
because, of course, it appears on TV here. Simple as that. We
wouldn't know any chefs, gardeners, DIY experts, antiques experts
or property developers either, were it not for popular TV series.
Spot the common theme.

I'm not quite sure I get your point. I guess I'm saying that
chess has a low (marketable product)/participant ratio.
As you seem to be aware, other activities which don't
seem any more inherently interesting than chess seem to find
a market. But chess refuses to.

A better point, IMO. And one which could do with broad and liberal
exploration.

...

It's
not a question that anyone can truly know the answer
to until it is tried. However, the lack of interest in
chess world played by those rules is fact, and it
seems reasonable to believe that there is a connection.

"Fact." Around here, the number of competitive chess
players is indeed seriously down from the "Fischer boom". OTOH,
it is still around twice the number from the '60s.

End of Fischer boom shoed USCF at about 50,000 members. Current number is
80,000 in US, although only 30,000 adults. Of those less than half play ANY
rated games per year, and of those only half play 10+ games, which would
make their rating 'current'.

Over that

same period, FIDE has expanded from a first-world club to include
practically every country in the world.

Fide's activities in Africa, for example, is down. Less support than
hertofore - so while there are greater number of coutries, extent is very
variable.

The ranks of GMs and IMs
have expanded from a handful of top names to a list several hundred
long. Junior chess events in the UK can attract thousands of
entrants, and dozens of coaches can earn a living. No interest?
Little publicity, perhaps.

I saw recently that an economic [!] correspondent said chess was second most
popular pasttime in UK, [after soccer].


Ah, but is it true? I used to be a habitual league player, but
nowadays play all my chess online. I may venture an OTB tournament in
the summer. Between 1993 and 1997 I used to play in about one every
month *as well as playing league matches*.

My story appears not to be uncommon, although it may be.

snips rest

Mark

"Dr A. N. Walker" wrote in message
...
In article ,
David Kane wrote:

Almost
every other popular activity has well-known stars
whose appeal extends well beyond that of expert
practitioners of the sport.
[...]
I'm not quite sure I get your point. I guess I'm saying that
chess has a low (marketable product)/participant ratio.

But that's not what you *were* saying. I merely pointed
you at a fairly long list of popular activities that do not have
"well-known stars", and then at the common theme of those that do
-- viz TV exposure.


But your list was full of things that are obviously nowhere
near as popular as chess, so it wasn't very persuasive.
In the US, a country with relatively little chess tradition,
the percentage of people with some interest in the game is
in double digit percentages. Ice dancing?

Put chess on reasonably prime-time TV, and
it will have stars; toss it to 4am on the minor channels and to
the back pages of the "quality" press, and it won't. Having a
"marketable product" is yet another quite different concept. ...

Your TV point was also unclear. Whether something makes it
on TV depends on how interesting it is. I'm proposing a way
that chess could be more interesting than it is.


As you seem to be aware, other activities which don't
seem any more inherently interesting than chess seem to find
a market. But chess refuses to.

... "Refuses"? Look, it's not a spectator sport. So
its "market" is elsewhe books, coaching, sponsorship, sets,
product placement, etc. You can get sponsorship not because
it's "exciting" but because it is associated with intellect,
because it's "educational", because it's something that people
who are disabled, poor, immigrant, young, old, ... can do,
because [broadly] it's *not* associated with crime, drugs, sex,
betting, .... There is quite a lot of money swilling around
chess, one way or another -- enough, world-wide, to support at
least several thousand full-time professionals [playing, coaching,
writing, directing, ...] -- it just isn't channelled into a
dozen or so big names who are always on our screens and thus
known to the general public. [But Kramnik isn't exactly in
poverty ....]

I'd love to hear your expert commentary of the last
round of the Tal Memorial tournament I previously
alluded to. If you can make that interesting, you should
be in the entertainment business.

I *am* in "the entertainment business"; and making a
game of chess interesting is trivial compared with making some
of the stuff I have to teach interesting. But no matter.

[...]
First, I don't understand your claim that computers are
one-dimensional.

You previously said you did!

They evaluate many different factors and then
combine them, just like humans. The difference is that
computers will give you a concrete explanation of how
they arrive at a move,

Not in any useful way. Note that the evaluation of a
*move* is merely the result of backing the evaluations of a
huge collection of *positions* up the game tree. The factors
that went into the static evaluation of *one* position perhaps
20 or 30 ply [with extensions] away from the *current* position
will usually make little sense. "You are 0.67 pawns up because
with what I think at the moment is best play you will arrive in
15 moves at a position where your opponent has doubled pawns.
If we play *your* move, you are only 0.65 pawns up, because in
12 moves we reach a position where you have won a piece for two
pawns and he has the two bishops."

But that's a precise description of how the computer arrives at its
answer. Humans say things like "castling king-side looked risky"

whereas we humans give vague
answers.

We could give the same answers, more-or-less ["I won
a piece for two pawns and some compensation in preference to
merely keeping an edge by doubling his pawns"]. That isn't
the real difference -- which is that the human has *reasons*
[possibly mistaken] for choosing one over the other, not just
a list of numbers from which to choose the largest.

But it doesn't change the underlying nature
of the process. I don't see any reason why computers couldn't
be programmed to do other things that humans do, like reduce
counterplay, [...].

Because you need to judge the counterplay as part of
the static evaluation [already almost a self-contradiction]

I disagree. A position that is mate has no counterplay.
A position that has a lopsided material balance with
low levels of material has little counterplay, etc.
What you'll find is that a computer evaluation function will
include many things that humans describe as relating to
counterplay. The big difference is that computers evaluate
*positions* but humans include external factors. I play
1.e4 against Kramnik and he plays 1. ...f6 and offers me
a draw, I take it. Not because of my objective evaluation
of the position on the chessboard, but because I evaluate
accepting a draw as being 0.5 points better than declining
the draw and playing something else.

and then back it up the tree, which mucks up the pruning that
enables the computer to go so deep [alpha-beta more-or-less
doubles the search depth, equivalent to an exponential increase
in the speed of your computer as it searches deeper].
are partial ways around this, but it's not easy. [This is
part of the reason why humans find it hard to adjust when the
psychology of a position changes -- if you thought you were
winning, and then find you aren't, it's very hard to take a
fresh objective look at the position.]

My reason for bringing up computers' low
draw rate is that it refutes the theory that high draw rates
are inherent result of a high Elo (i.e. the perfectly played
game is a draw and humans play perfectly enough) It
suggests that the human draw rate stems from factors
external to the game [for example, overcounting draws].

Or from factors related to differences in the way
humans and computers play chess ....

Yes, but what are those specifically? Humans see
that draws help their tournament situation and play
drawish moves.

I'm just saying let's experiment with modifying the
incentives to produce more interesting competitions.
Well, that's fine. But you do your case no good by your
use of language. *I* don't feel that my own draws are not real
contests,
I never claimed that your games weren't real contests. Only
that everybody's games could be more contested if
the external incentives were different.

So you claim that I should feel that, though my own
games are real, other people are playing less competitively?
Even if true, not good politics.

I'm suggesting that in general people (including you) play
less competitively than they could but never
claimed that that makes your games "unreal".
You remind of the people who claim that
advertising doesn't work on them. But we
know that in aggregate advertising does work. Are
you claiming that your play is completely uninfluenced
by the knowledge that a drawn result will score you
points? If you played in a tournament scored with
BAP, you'd play *exactly* as you would without it?
Fact is, when strong players did play in a BAP
tournament, their behavior *was* different. They
played more interesting chess!


nor that chess is boring because some games are drawn,
I definitely never claimed that. I've been talking all along
about *excessive* draw rates.

But you divert all the time from "excessive" [if that
is true] draws to lack of competitiveness, boredom and other
consequences for popularity, marketability, ....

It's not a diversion, but a progression. Chess uses a
scoring system that makes the risk/reward of playing for a win
very poor. At the GM level, this leads to lots of draws,
and lots of tournaments with prominent strategies based
on draws. These are completely at odds with what human
beings normally consider interesting and dramatic, so few,
even the millions who enjoy playing chess, pay attention
to the games.


nor that drawing two games against a GM opponent are less of an
achievement [for me!] than a win and a loss,
Mathematically, it *is* less of an achievement
because GMs don't lose very often. How you feel
about the achievement is, of course, up to you.

But I am a good enough player to draw, occasionally,
against much stronger players, even GMs, on *merit*, without
their making a manifest error. I am not good enough to beat
a GM unless he does something really daft [which, of course,
even GMs do on occasion]. [The last time I beat a player
rated over 2400, it was a league match, and my opponent was
the only person in the room who did not know Wurzberger's
trap. Both my team and his were falling about with laughter,
and I was pleased enough to win, but didn't feel it was that
much of an achievement.]

I have similar feelings from my own games, though I'm a class
player. But it would never occur to me to argue that pride in
my draws with masters would be a valid reason to favor a
system that *obviously* harms the GM game. And it *is*
obvious - just look at the last round of the Tal Memorial
tournament I referred to.


nor that limited
Serious Money in chess is proof that something is Wrong,
Proof, possibly not. But it sure has to be evidence of something.
Why deny it?

I don't deny it. But evidence of something is not
evidence that something is Wrong.

nor
that the current Laws are outmoded.
I believe that the *scoring system* is outmoded.

Fine; you're entitled to your belief, and to try
to convince the rest of us. But the rest of us are more
likely to be convinced by rational argument that by being
told that we -- who have *chosen* to be interested in chess
as it is -- are playing by a boring, outmoded system.

I *have* wondered whether many people are attracted
to chess precisely *because* draws feature so prominently
in the game. Perhaps they fufill some psychological yearning
for "fairness". In that case, giving chess the features
we usually associate with competitions (winners, drama)
might make the game more popular in general, but less
popular to its current adherents. I think that is why so
many of the anti-anti-draw posts tend to the visceral
rather than the rational. People *like* their draws and
really don't want to analyze the consequences of their
preference.

In article ,
David Kane wrote:
I'm not quite sure I get your point. I guess I'm saying that
chess has a low (marketable product)/participant ratio.
But that's not what you *were* saying. I merely pointed
you at a fairly long list of popular activities that do not have
"well-known stars", and then at the common theme of those that do
-- viz TV exposure.
But your list was full of things that are obviously nowhere
near as popular as chess, so it wasn't very persuasive.

You're moving goalposts again! Plenty of activities
that are equally nowhere near as popular as chess *do* have
stars; plenty don't. Plenty are [or are not] spectator
activities, are [are not] on TV, do [do not] attract lots
of money, and so on. Little to do with draws/results, or
we wouldn't have star chefs, property developers, etc.

In the US, a country with relatively little chess tradition,
the percentage of people with some interest in the game is
in double digit percentages. Ice dancing?

What do you mean by "some interest"? Ice dancing was
on prime-time TV in the UK recently as a "celebrity reality"
show -- Torvill and Dean + assorted experts/celebs learning to
skate and dance + voting -- and it was massively popular. Lots
of people are participating in ice activities of some sort --
our "Ice Arena" is throbbing with amateurs and professionals --
far more so than are participating in chess. And there are large
crowds at the ice hockey [which gets no real TV coverage, and so
has no stars here]. Lots of people "know the moves" of chess;
only a few play competitively [less than 0.1% in the UK]. The
mix between stars, professionals, competitors, participants,
spectators and those with "some interest" is different in every
activity, be it a sport, a pastime, a game, a hobby or just an
activity, and to isolate draws as a main, or even a significant,
reason for the particular mix in chess seems to me OTT.

Not in any useful way. Note that the evaluation of a
*move* is merely the result of backing the evaluations of a
huge collection of *positions* up the game tree. [...]
But that's a precise description of how the computer arrives at its
answer.

Quite; and it had *nothing* to do with "moves" or "plans",
merely with features of positions.

Humans say things like "castling king-side looked risky"

And in one phrase, you have instead talked about a move and
by inference a plan. I may play a *move* with the *purpose* of
giving my opponent doubled pawns or of attacking his castled king;
the computer looks at the *position* 30 ply [or whatever] hence and
*notices* the doubled pawn or the pieces surrounding his king. It's
a quite different process.

Because you need to judge the counterplay as part of
the static evaluation [already almost a self-contradiction]
I disagree. A position that is mate has no counterplay.
A position that has a lopsided material balance with
low levels of material has little counterplay, etc.

And there you are talking about terminal or leaf positions
where the counterplay is easily determined. A *static* evaluation
has no way to decide whether [eg] a passed pawn or an attack is
really dangerous [that requires further *dynamic* (tree-searching)
evaluation] beyond counting features.

What you'll find is that a computer evaluation function will
include many things that humans describe as relating to
counterplay. The big difference is that computers evaluate
*positions* but humans include external factors.

The big difference is that computers can't [easily]
back that counterplay up the tree. Once you're away from the
leaves, all the computer sees is single numbers. No "buts".
At the *leaf* it may see "I'm a piece up for two pawns, but
he gets some play" -- but that turns into 3.0-2.0-0.5 = 0.5
pawns advantage and is indistinguishable from any other 0.5
pawn advantage, such as "he has doubled pawns". The computer
*could* pass further info back up the tree, but if so it would
prevent much of the pruning, and dramatically reduce the amount
of searching possible. Some "+0.5" positions are safe draws
with some winning chances, others are "it's won or lost, and
there's a slightly better chance of winning than losing", and
the computer has no idea what its analysis will show when that
position -- or one of the others in that part of the tree --
is actually reached and can be analysed further.

I play
1.e4 against Kramnik and he plays 1. ...f6 and offers me
a draw, I take it. [...]

OK, but that's another matter entirely [and not one which
is likely to happen to either of us in reality].

[...] Are
you claiming that your play is completely uninfluenced
by the knowledge that a drawn result will score you
points?

No. But I don't think the influence is entirely the
way you propose. Reducing points for a draw will certainly
cause people with the better of a draw to take more risks to
try to win. But it equally causes people with the worse of
a draw to take less care not to lose. Neither of these will
enhance the *quality* of play; the first may well enhance
the *interest*, but the second diminishes it. As there are
exactly as many players on the worse side of a draw as on
the better, the overall balance is unclear.

Two other points while I'm here. As soon as the score
per game is not zero-sum, you have provided a clear incentive
for result fixing. "If this game, as played fair and square,
turns out to be clearly drawn, then instead of agreeing that,
we will toss a coin to decide who makes the blunder." Also,
even if the game score is not zero-sum, the rating points
still have to be. You could finish with the worst player in
the tournament by performance rating being the champion.

I *have* wondered whether many people are attracted
to chess precisely *because* draws feature so prominently
in the game.

Seems very unlikely, as draws are rare in the games of
beginners and of low-rated players generally. So it's long
after people become "hooked" that they realise about draws.

--
Andy Walker, School of MathSci., Univ. of Nott'm, UK.

"Dr A. N. Walker" wrote in message
...
In article ,

I play
1.e4 against Kramnik and he plays 1. ...f6 and offers me
a draw, I take it. [...]

OK, but that's another matter entirely [and not one which
is likely to happen to either of us in reality].

It's not another matter. It's the matter I have been talking
about all along. Chess' scoring influences the play, in
a very detrimental way.

And while that exact scenario may be unlikely (who knows,
perhaps I will face Kramnik while he is trying to win the
$1,000,000 prize offered by Mr. Houlsby for being the
first person to draw 10,000 consecutive chess games)
it's the kind of influence that is *always* present.

[...] Are
you claiming that your play is completely uninfluenced
by the knowledge that a drawn result will score you
points?

No. But I don't think the influence is entirely the
way you propose. Reducing points for a draw will certainly
cause people with the better of a draw to take more risks to
try to win. But it equally causes people with the worse of
a draw to take less care not to lose. Neither of these will
enhance the *quality* of play; the first may well enhance
the *interest*, but the second diminishes it. As there are
exactly as many players on the worse side of a draw as on
the better, the overall balance is unclear.

They start the game knowing they will have a battle in
front of them. The option of an early truce makes no sense
now. They play Sicilians instead of Caro-Kans. They seek
double-edged positions rather than simplifying to an
equal endgame they know how to draw. In short, they
play chess.

One could debate forever exactly which scoring incentives
produce the "best" game, and different people could have
different opinions, even once the connection between
the incentives and the play are empirically established.
I don't see that there is any credible debate over
whether 1867-scoring has glaring defects.

Two other points while I'm here. As soon as the score
per game is not zero-sum, you have provided a clear incentive
for result fixing. "If this game, as played fair and square,
turns out to be clearly drawn, then instead of agreeing that,
we will toss a coin to decide who makes the blunder." Also,
even if the game score is not zero-sum, the rating points
still have to be. You could finish with the worst player in
the tournament by performance rating being the champion.

This straw argument is frequently made. Chess,
as played now, is already not zero sum because
prizes are awarded for place finish, not proportional
to points. So players can currently make more
money by rigging their results. What happens when
draws are discounted (BTW that is not the only
way to increase the contestedness of games, but it
is an often discussed one) is that the two players
will collude in a good way by going into
complex, double-edged lines, which would have a higher
probability of rewarding the player who plays better with
a win. Of course, at times they won't be successful and
they'll end up with a hard fought draw. I have never
heard anyone say that is a bad thing.

Ratings can also be made more accurate by adjusting
the rating algorithm to account for WLD statistics.
Clyde Ballard, of BAP fame, has done some interesting
analysis. In any case, ELO's methods can be
readily adapted to alternate scoring systems.
[As an aside, one proposal that I have thought
has some merit is that only decisive
games be rated. This makes sense from the
standpoint that many drawn results are not
seriously contested chess games but rather
pseudo-games in response to external factors.
It also eliminates one of the incentives for the
quick last round draw "I'll give you some of
my rating points if you don't compete for
my prize money"]

In article ,
David Kane wrote:
I play
1.e4 against Kramnik and he plays 1. ...f6 and offers me
a draw, I take it. [...]
OK, but that's another matter entirely [and not one which
is likely to happen to either of us in reality].
It's not another matter. It's the matter I have been talking
about all along.

No it isn't. You've been talking about popularity,
stars, marketing, the relation between computer and human
play, and many other things. AAMOF, nothing remotely like
your Kramnik scenario has *ever* happened to me, in over 50
years of playing chess. I *have* known it to happen, in
cases where players who were thoroughly fed up with their
performances in a tournament just wanted to call it a day
in the last round; which will happen in any scoring system.

No-one is defending the 15-move GM draw; even the
GMs concerned mostly wouldn't do so in chessic terms, but
rather in terms of their mental/physical state. If you want
to stop that being a factor in the popularity of chess, you
need to change that state, not meddle with scores. [And it's
worth noting that the only chess that routinely gets into the
general news items is the world championship match, for which
no amount of tinkering with scores -- short of awarding more
for wins with black -- is going to affect strategy.]

[...] As there are
exactly as many players on the worse side of a draw as on
the better, the overall balance is unclear.
They start the game knowing they will have a battle in
front of them. The option of an early truce makes no sense
now. They play Sicilians instead of Caro-Kans.

Whoa! I'm quite sure that when IMs and GMs play the
Caro-Kann [or the Pirc, French, etc] against me rather than
the Sicilian, it's not with a view to an early truce. Most
GMs playing a weaker player want to reduce the variance --
they would much rather have an edge which they know they will
be able to turn into the full point in the ending than an
unfathomable position where they just might find themselves
on the wrong side of a combination.

You have also given players a further incentive to
play solidly, on the grounds that an opponent in a must-win
situation will have to take unsound risks.

It really is not as simple as you seem to think. You
will certainly reduce the incidence of draws if you reduce
their value; but this is not the same as increasing either
the interest or the quality of the games.

[...]
I don't see that there is any credible debate over
whether 1867-scoring has glaring defects.

I'm afraid that is your problem. Until you do see that,
you are simply shouting at us.

This straw argument is frequently made. Chess,
as played now, is already not zero sum because
prizes are awarded for place finish, not proportional
to points.

But neither the points nor the prizes are changed in
total by this. I could perhaps collude with you to increase
our total prize -- but this would be at the expense of someone
else who has been deprived. I could also already perhaps get
some title by collusion. You are proposing nothing to reduce
these possibilities -- which already happen, and are regarded
by most of us as fraud. It's not a straw argument to point
our that you are adding a *further* collusion possibility --
esp not when below you add a potential rating fraud to it.

Ratings can also be made more accurate by adjusting
the rating algorithm to account for WLD statistics. [...]

They could certainly take some account of colour and
of variance. But they are not going to be more accurate if
they cease to be [broadly] zero-sum [there are drifts caused
by some people playing more than others]; nor if they fail
to recognise that for me to draw with a GM is a considerable
achievement [and correspondingly a failure for him].

--
Andy Walker, School of MathSci., Univ. of Nott'm, UK.

"Dr A. N. Walker" wrote in message
...
In article ,
David Kane wrote:
I play
1.e4 against Kramnik and he plays 1. ...f6 and offers me
a draw, I take it. [...]
OK, but that's another matter entirely [and not one which
is likely to happen to either of us in reality].
It's not another matter. It's the matter I have been talking
about all along.

No it isn't. You've been talking about popularity,
stars, marketing, the relation between computer and human
play, and many other things. AAMOF, nothing remotely like
your Kramnik scenario has *ever* happened to me, in over 50
years of playing chess. I *have* known it to happen, in
cases where players who were thoroughly fed up with their
performances in a tournament just wanted to call it a day
in the last round; which will happen in any scoring system.

The central point has been the same all along. The various
sideshows have cropped up largely in response to your
introducing them.

No-one is defending the 15-move GM draw; even the
GMs concerned mostly wouldn't do so in chessic terms, but
rather in terms of their mental/physical state.

In fact, *you* have defended them.

You offered being "tired", "risk averse" and "Get me
out of here." as being valid reasons for the 2700
players in the Tal Memorial to play perfunctory
draws in their last round. Those very well could
be the reasons the players had, and, personally,
I don't fault them for making
what was probably a rational decision.

The point is that we should learn from it. First is
understanding that their behavior was rewarded
by the *external incentive structure*. Second, this
sort of behavior is abnormal in just about
every other competitive activity, esp. those
with any following. If France and Italy are "tired" or
"risk averse", can they decide not to play the
World Cup finals??




to stop that being a factor in the popularity of chess, you
need to change that state, not meddle with scores.

Said without a shred of evidence. In the one GM BAP
tournament, there were several players in contention
going into the last round and the result was 100% decisive
games, not unplayed draws. That's the way competitions
should be.

Moreover, tired players, *placed in a truly competitive
situation* should be less draw prone, not more draw prone.
Most competitive endeavors do not give their contestants
an option of napping in the competition, and common
sense tells us that neither should chess.

[And it's
worth noting that the only chess that routinely gets into the
general news items is the world championship match, for which
no amount of tinkering with scores -- short of awarding more
for wins with black -- is going to affect strategy.]

This is a valid point which I have made myself. But it doesn't
have anything to do with the general proposition of whether
competitive chess can be made more interesting by modifying
the perverse incentives under which it is currently played.


snipped

It really is not as simple as you seem to think. You
will certainly reduce the incidence of draws if you reduce
their value; but this is not the same as increasing either
the interest or the quality of the games.


As I've said, reducing the number of draws isn't
the real objective so much as increasing the contestedness
of each game. No one denies that there will still
be a natural draw rate - blitz games between E players
on the internet sometimes end in draws.

The fact that there are quite a number of barely
contested games in major chess events is
proof that there is something wrong with the
incentive structure in chess.

[...]
I don't see that there is any credible debate over
whether 1867-scoring has glaring defects.

I'm afraid that is your problem. Until you do see that,
you are simply shouting at us.

I've never seen *your* non-evasive (please no
bizarre fishing analogies or tangential semantic
arguments about what is meant by "counterplay")
reply to my argument that:

Chess uses a scoring system that makes the risk/reward
of playing for a win at the GM level very poor. This leads
to lots of draws, and lots of tournaments with prominent
strategies based on drawing and playing to draw. These
are at odds with what human beings normally consider
interesting and dramatic, so few, even among the millions
who enjoy playing chess, pay attention to the games.

On 5 Apr, 20:40, "David Kane" wrote:
"Dr A. N. Walker" wrote in ...





In article ,
David Kane wrote:
I play
1.e4 against Kramnik and he plays 1. ...f6 and offers me
a draw, I take it. [...]
OK, but that's another matter entirely [and not one which
is likely to happen to either of us in reality].
It's not another matter. It's the matter I have been talking
about all along.

No it isn't. You've been talking about popularity,
stars, marketing, the relation between computer and human
play, and many other things. AAMOF, nothing remotely like
your Kramnik scenario has *ever* happened to me, in over 50
years of playing chess. I *have* known it to happen, in
cases where players who were thoroughly fed up with their
performances in a tournament just wanted to call it a day
in the last round; which will happen in any scoring system.

The central point has been the same all along. The various
sideshows have cropped up largely in response to your
introducing them.

No-one is defending the 15-move GM draw; even the
GMs concerned mostly wouldn't do so in chessic terms, but
rather in terms of their mental/physical state.

In fact, *you* have defended them.


No, he hasn't, and neither have I. You really need to learn to read.

You offered being "tired", "risk averse" and "Get me
out of here." as being valid reasons for the 2700
players in the Tal Memorial to play perfunctory
draws in their last round. Those very well could
be the reasons the players had, and, personally,
I don't fault them for making
what was probably a rational decision.


That's good.

The point is that we should learn from it.

That's the royal "we" I take it?

First is
understanding that their behavior was rewarded
by the *external incentive structure*.

Nonsense.

Second, this
sort of behavior is abnormal in just about
every other competitive activity, esp. those
with any following.

Nonsense. It happens in snooker. When England's national soccer team
plays, it sometimes happens in soccer too.

If France and Italy are "tired" or
"risk averse", can they decide not to play the
World Cup finals??


This is a stupid question from a stupid individual. Read the archive.

to stop that being a factor in the popularity of chess, you
need to change that state, not meddle with scores.

Said without a shred of evidence.

READ AN INFORMATOR, YOU ****ING MORON.

snips more Kane trolling bull****

"David Kane" wrote in message
. ..

"Dr A. N. Walker" wrote in message
Not in any useful way. Note that the evaluation of a
*move* is merely the result of backing the evaluations of a
huge collection of *positions* up the game tree. The factors
that went into the static evaluation of *one* position perhaps
20 or 30 ply [with extensions] away from the *current* position
will usually make little sense. "You are 0.67 pawns up because
with what I think at the moment is best play you will arrive in
15 moves at a position where your opponent has doubled pawns.
If we play *your* move, you are only 0.65 pawns up, because in
12 moves we reach a position where you have won a piece for two
pawns and he has the two bishops."

But that's a precise description of how the computer arrives at its
answer. Humans say things like "castling king-side looked risky"

This is an interesting aspect of both human and computer evaluation. I am
not sure that our quantitative measures are even uniformly agreed upon - how
many point is a bishop worth, eg, if a Kt is nominally 3.0. Fischer said
3.25. Spassky 'made a living' from BxN. If you only have one bishop, is it
still worth 3.25, or is it now worth 2.75? Or does it depend if the other
guy has 2 bishops?

it get more interesting when you have to evaluate 2 bishops strategically -
in human terms we think they are 'good' in an open position, but sometimes
still play for two bishops in an early closed position, since we anticipate
in 15 or more moves the position will open up

but this is precisely the sort of evaluation that computers do not make,
since even 'deep' evaluations are often just 12/13 moves

I anticipate a future conversation here in respect of the new MAMS title,
which points this out as a specific short-coming of computer evaluation -
and even a relatively easy one compared with, say, sacking a pawn in a
gambit and gaining the initiative where the silly machine keeps evaluating
itself as +1 or better without compensating for the lost initiative

if you read Russian chess bulletins or commentaries, often the commentator
goes nutz@ during a game, and enthusiastically applauds who ever has seized
the initiative, almost as if = 'game over', but certainly that chances now
reside with the player with the initiative since they confers control

I think the subject is complex [that is, a combination of several factors -
in fact, of several dynamic factors] and this is a weakness of computer
programs operating from fixed and pre-programmed evaluations

The problem must be when to re-evaluate the worth of the pieces, ie, when
are 2 bishops worth 3.25 each? when less, and when to switch evaluations?

On the human side, there appears to be better process to evaluate /dynamic/
aspects of the game, ie, "castling king-side looked risky" or 'yes, I'm
going to play this gambit' although those things are, while better perceived
as delivering a better quality of game, are more difficult to quantify.

Phil Innes